Due to the poor predictive validity of most infant assessment procedures, research on the developmental implications of alcohol-related central nervous system (CNS) deficits has required expensive longterm follow-up studies. Recently, however, infant development researchers have devised new assessment procedures focusing on particular aspects of cognitive processing which have demonstrated markedly better predictive validity than traditional infant tests. These newer tests also appear to be more sensitive to subtle CNS damage possibly associated with moderate levels of maternal drinking during pregnancy and, therefore, are particularly well-suited for investigating thresholds of toxic effects. All patients seen in a prenatal clinic during a 22-month period will be screened for alcohol intake at their first clinic visit. A sample of approximately 530 subjects will be selected to include all heavier drinkers and known proportions of lighter drinkers and abstainers. Birthweight, gestational age, major congenital anomalies, and dysmorphology will be ascertained at birth. The infants will be tested at home on Fagan's visual recognition memory test, cross-modal transfer, and object permanence at the appropriate postnatal ages and on the Bayley Scales of Infant Development at 12 1/2 months. It is hypothesized that the newer, more focused cognitive tasks will be more sensitive to the effects of prenatal alcohol exposure than the Bayley Scales and that cognitive deficits will appear at thresholds below those associated with anatomic abnormalities, such as dysmorphology and intrauterine growth retardation. It is further hypothesized that effects on infant cognitive functioning will not be mediated by socio-environmental factors, such as quality of home environment and postnatal maternal drinking. The combination of cognitive measures providing the most efficient assessment of the teratogenic effects of prenatal alcohol exposure will be computed, and the influence of timing (i.e., trimester of exposure) and pattern of maternal drinking (steady vs. binge) will be evaluated. Because research on alcohol-related birth deficits is by necessity correlational, data will also be obtained on 20 prenatal control variables, including demographic background, obstetrical risk, and exposure to other drugs, adn 16 postnatal control variables, including parental involvement with the infant and infant temperament. The impact of covariates of prenatal alcohol exposure will be controlled statistically in the data analyses.